what we do

Our focus is on the development of proprietary geodesic models that characterize the behavior of leveraged long/short ETFs in pairs trades using first principles of Hamilton-Lagrange-Euler mechanics. Because of tracking errors & daily compounding phenomena, equal weightings of leveraged ETFs in a pairs trade are virtually never 50-50. Daily data & graphics will show subscribers where neutral pair weightings have moved along the path of the pairs geodesic thereby providing multiple market direction & re-balance indicators.

blog archive

Saturday, April 11, 2009

SDS-SSO Subscriber Alert for 2009-04-11

Dear Subscriber,

greetings all,

i'm making some significant (good) progress with re-characterizing the PID control portion of the model. the control aspect of the model is the mechanism to tell you when it's time to re-balance based on boundary conditions i set. %change(m) will still play a role in the control loop but with additional components now. i'm still backtesting the controls and will settle on which paramerters to start with going into monday. this will be reflected in the analysis i need to send you with last thursday's update.

in the meantime, here's a breif summary of what this is all about. the control loop looks at the benchmark pair-value and computes an error signal that is pair-value minus 100, where 100 is the ideal neutral value of the pair. when the control loop senses this error is outside limits i set, that becomes the signal to re-balance the benchmark. next i look at the aggressive performance scenario and i go through a check list to see if it needs to be re-balanced and which way. the process to determine this requires more evaluation. i have some initial criteria that give good performance but they will need more back-testing and, frankly, more trading days because the ETFs have only been on the market for maybe 2 years max.

this prompted the figurative "light bulb" to light up over my head with an idea to possibly take aggressive optimization to another level, perhaps using neural net algos, fuzzy logic or genetic evolvers. but these subjects might be over-kill or intangible for general use. i want the model to be as easy as possible to use for most anyone knowledgable enough to know the basics of the ETF asset class, particularly levered ETFs. an elementry optimization process that works 90 - 95% of the time is right in the pocket.

more coming up shortly!
best regards,
mike james

Managing Member
Equity Informatics, LLC
phone:302-220-3864

No comments:

Disclaimer

Equity Informatics is a developer and service provider of proprietary financial equity pricing models & trading methods. The company familiarizes subscribers with the basic thesis of our models, provides subscribers with daily neutral pair weightings and methodologies on how to use the data as intended. subscribers shall not share any information obtained from equity informatics with any other party. use of these services are granted only to and intended for the benefit of the subscriber. Equity Informatics does not offer the sale of equities nor do our trading models constitute trading advise. It is incumbent on potential clients to perform due diligence and seek a professional financial adviser to help you determine whether subscribing to the company's services are suitable for your financial situation and level of risk. No guarentees of performance are expressly or implicitly offered nor does Equity Informatics guarantee the accuracy of market information used to provide model data to our client. equity informatics does not assume responsibility for lost principal, lost gains or tax consequences.

Copyright (c) 2009 Equity Informatics, LLC. All Rights Reserved.